A Multiple Features Image Tracking Algorithm

摘要：

In Mean Shift algorithm, the features of the tracked target and the image matching similarity criterion have great influence on the result of tracking. A new algorithm of target tracking is proposed. The algorithm combine local binary pattern and color information to form a new feature CL, which tracks target by using a method of centroid iteration based on maximum posterior probability. Thanks to the simplification of the LBP, the CL has higher differentiation ability and lower computational complexity. Experimental results show that the new algorithm have significantly improved the tracking performance, in comparison with original Mean Shift algorithm. In complex background, the algorithm can track the target robustly.

关键词： object tracking Local Binary Patterns maximum posterior probability multiple features fusion Mean Shift

作者: Wenhua Guo Zuren Feng Shuai Wang Qin Nie

作者单位: Systems Engineering Institute, Slate Key Laboratory for Manufacturing Systems Engineering, Xi an Jiaotong University, Xi an 710049

会议类型: 国际会议

会议名称: 2012 Fifth International Symposium on Computational Intelligence and Design 第五届计算智能与设计国际会议 ISCID 2012

会议地点: 杭州

会议语种:英文

页码: 655-658

在线出版日期: 2012-10-28（万方平台首次上网日期，不代表论文的发表时间）